Question on storage

WebODM
#1

Good evening,

I had a question regarding where the data would be stored.

Here is current setup

WebODM on a PC#1 for GUI interface.
ClusterODM setup on another PC#2 for routing
Node1, Node2 and Node3 setup on server for processing power.

When you create a project through the dashboard and upload the data, where does that data go and store? Does it send and store to PC#1 or does it route that data and store it to whichever node gets selected by Cluster?

If it does store on PC#1 and let’s say it sends it to Node1. Is it reading the data over the network off of PC #1 or does PC #1 send that data and it reads/writes/renders off of Node 1?

Just trying to determine where I should focus my attention on storage/speed etc.

1 Like
#2

WebODM stores a copy of the images.
ClusterODM does not store the data long term, it forwards it to the nodes, although it will store it temporarily until the data has been forwarded.
Nodes store the images for 2 days by default (you can lower that via the node’s settings if needed).

1 Like
#3

So just want to clarify. I upload the data to WebODM (PC #1) This will store a copy of the data, the same data will also send to whichever node takes the project (Server) I just was curious is it reading and rendering the data from the data off the Node or off of WebODM when processing.

Also, one last question, is there an option when setting the parameters to run a job, to skip plant health and not make that an option?

Thank you

1 Like
#4

That is calculated on-the-fly at time of view/request in the View Map/2D View function, so it isn’t costing you anything during processing.

#5

Thank you. Just wanted to confirm again.

When working with with WebODM and NodeODM through Cluster, when a project gets added and it forwards it to one of the NodeODM, does it read/render off of WebODM setup or NodeODM setup?

Basically, on my WebODM server, it’s running HDD because I can get more storage, on the NodeODM server, it’s running off of SSDs. For performance, I was trying to figure if it’s reading and rendering off of the HDD or the SSDs. I know Pierotofy had mentioned it sends a copy on WebODM, just didn’t know which device it’s reading and rendering off of. I’m debating if I should just put SSDs in the WebODM server, but SSD’s just don’t have as much storage for the price point.

1 Like
#6

I believe it is going to be done via the WebODM instance, but it really shouldn’t need to touch the disk terribly much aside from reading the COG.

1 Like
#7

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.